[FFmpeg-devel] [Patch] QT RLE encoder, bis

[Alexis Ballier]

> and after this is applied, maybe you want to add rate distortion optimal
> encoding support
> that is minimize number_of_bits*constant + sum_of_squared_differences instead
> of just number_of_bits

I'm not sure I understand it correctly, what are the differences that
can be taken into account ?
The difference of the number of bits in a line ? Of the ones of a frame ?
Or do you mean a more complex algorithm that would "approximate" the
pixels (in order to be able to use "skip" or "repeat" more often), and
thus the differences would be between the original and the encoded
pixel ?

The first two would increase the size of the encoded video. The last
one could/will achieve better compression but would make it lossy.

Alexis

Ps: And many thanks for your help improving & cleaning the code